Suppr超能文献

用于数字乳腺摄影质量控制的调制传递函数(MTF)测量的验证

Validation of MTF measurement for digital mammography quality control.

作者信息

Carton Ann-Katherine, Vandenbroucke Dirk, Struye Luc, Maidment Andrew D A, Kao Yen-Hong, Albert Michael, Bosmans Hilde, Marchal Guy

机构信息

University Hospitals of Leuven, Herestraat 49, Leuven, Brabant 3000, Belgium.

出版信息

Med Phys. 2005 Jun;32(6):1684-95. doi: 10.1118/1.1921667.

Abstract

The modulation transfer function (MTF) describes the spatial resolution properties of imaging systems. In this work, the accuracy of our implementation of the edge method for calculating the presampled MTF was examined. Synthetic edge images with known MTF were used as gold standards for determining the robustness of the edge method. These images simulated realistic data from clinical digital mammography systems, and contained intrinsic system factors that could affect the MTF accuracy, such as noise, scatter, and flat-field nonuniformities. Our algorithm is not influenced by detector dose variations for MTF accuracy up to 1/2 the sampling frequency. We investigated several methods for noise reduction, including truncating the supersampled line spread function (LSF), windowing the LSF, applying a local exponential fit to the LSF, and applying a monotonic constraint to the supersampled edge spread function. Only the monotonic constraint did not introduce a systematic error; the other methods could result in MTF underestimation. Overall, our edge method consistently computed MTFs which were in good agreement with the true MTF. The edge method was then applied to images from a commercial storage-phosphor based digital mammography system. The calculated MTF was affected by the size (sides of 2.5, 5, or 10 cm) and the composition (lead or tungsten) of the edge device. However, the effects on the MTF were observed only with regard to the low frequency drop (LFD). Scatter nonuniformity was dependent on edge size, and could lead to slight underestimation of LFD. Nevertheless, this negative effect could be minimized by using an edge of 5 cm or larger. An edge composed of lead is susceptible to L-fluorescence, which causes overestimation of the LFD. The results of this work are intended to underline the need for clear guidelines if the MTF is to be given a more crucial role in acceptance tests and routine assessment of digital mammography systems: the MTF algorithm and edge object test tool need to be publicly validated.

摘要

调制传递函数(MTF)描述了成像系统的空间分辨率特性。在本研究中,我们检验了用于计算预采样MTF的边缘法实现的准确性。将具有已知MTF的合成边缘图像用作确定边缘法稳健性的金标准。这些图像模拟了临床数字乳腺摄影系统的真实数据,并包含可能影响MTF准确性的内在系统因素,如噪声、散射和平场不均匀性。我们的算法在MTF准确性达到采样频率的1/2时不受探测器剂量变化的影响。我们研究了几种降噪方法,包括截断超采样线扩展函数(LSF)、对LSF加窗、对LSF进行局部指数拟合以及对超采样边缘扩展函数施加单调约束。只有单调约束没有引入系统误差;其他方法可能导致MTF低估。总体而言,我们的边缘法始终能计算出与真实MTF高度一致的MTF。然后将边缘法应用于基于商用存储磷光体的数字乳腺摄影系统的图像。计算得到的MTF受边缘装置的尺寸(边长为2.5、5或10厘米)和组成(铅或钨)影响。然而,对MTF的影响仅在低频下降(LFD)方面被观察到。散射不均匀性取决于边缘尺寸,可能导致LFD略有低估。尽管如此,通过使用5厘米或更大尺寸的边缘,这种负面影响可以最小化。由铅组成的边缘易受L荧光影响,这会导致LFD高估。如果MTF要在数字乳腺摄影系统的验收测试和常规评估中发挥更关键的作用,本研究结果旨在强调明确指导方针的必要性:MTF算法和边缘物体测试工具需要进行公开验证。

文献AI研究员

20分钟写一篇综述,助力文献阅读效率提升50倍。

立即体验

用中文搜PubMed

大模型驱动的PubMed中文搜索引擎

马上搜索

文档翻译

学术文献翻译模型,支持多种主流文档格式。

立即体验